[PATCH v3] mm: vmalloc: WARN_ON if mapping size is not PAGE_SIZE aligned

[PATCH v3] mm: vmalloc: WARN_ON if mapping size is not PAGE_SIZE aligned

[Yadong Qi]

In mm/vmalloc.c, the function vmap_pte_range() assumes that the
mapping size is aligned to PAGE_SIZE. If this assumption is
violated, the loop will become infinite because the termination
condition (`addr != end`) will never be met. This can lead to
overwriting other VA ranges and/or random pages physically follow
the page table.

It's the caller's responsibility to ensure that the mapping size
is aligned to PAGE_SIZE. However, the memory corruption is hard
to root cause. To identify the programming error in the caller
easier, check whether the mapping size is PAGE_SIZE aligned with
WARN_ON().

Signed-off-by: Yadong Qi <yadong.qi@linux.alibaba.com>
Reviewed-by: Huang Ying <ying.huang@linux.alibaba.com>
---
v2 -> v3:
  * change error code from ENOMEM to EINVAL
  * modify callers of vmap_pte_range to handle return code
v1 -> v2:
  * Use WARN_ON instead of BUG_ON
---
 mm/vmalloc.c | 29 ++++++++++++++++++-----------
 1 file changed, 18 insertions(+), 11 deletions(-)

diff --git a/mm/vmalloc.c b/mm/vmalloc.c
index 5edd536ba9d2..1fa52f203795 100644
--- a/mm/vmalloc.c
+++ b/mm/vmalloc.c
@@ -100,6 +100,9 @@ static int vmap_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end,
 	struct page *page;
 	unsigned long size = PAGE_SIZE;
 
+	if (WARN_ON(!PAGE_ALIGNED(end - addr)))
+		return -EINVAL;
+
 	pfn = phys_addr >> PAGE_SHIFT;
 	pte = pte_alloc_kernel_track(pmd, addr, mask);
 	if (!pte)
@@ -167,6 +170,7 @@ static int vmap_pmd_range(pud_t *pud, unsigned long addr, unsigned long end,
 {
 	pmd_t *pmd;
 	unsigned long next;
+	int err;
 
 	pmd = pmd_alloc_track(&init_mm, pud, addr, mask);
 	if (!pmd)
@@ -180,10 +184,11 @@ static int vmap_pmd_range(pud_t *pud, unsigned long addr, unsigned long end,
 			continue;
 		}
 
-		if (vmap_pte_range(pmd, addr, next, phys_addr, prot, max_page_shift, mask))
-			return -ENOMEM;
+		err = vmap_pte_range(pmd, addr, next, phys_addr, prot, max_page_shift, mask);
+		if (err)
+			break;
 	} while (pmd++, phys_addr += (next - addr), addr = next, addr != end);
-	return 0;
+	return err;
 }
 
 static int vmap_try_huge_pud(pud_t *pud, unsigned long addr, unsigned long end,
@@ -217,6 +222,7 @@ static int vmap_pud_range(p4d_t *p4d, unsigned long addr, unsigned long end,
 {
 	pud_t *pud;
 	unsigned long next;
+	int err;
 
 	pud = pud_alloc_track(&init_mm, p4d, addr, mask);
 	if (!pud)
@@ -230,11 +236,11 @@ static int vmap_pud_range(p4d_t *p4d, unsigned long addr, unsigned long end,
 			continue;
 		}
 
-		if (vmap_pmd_range(pud, addr, next, phys_addr, prot,
-					max_page_shift, mask))
-			return -ENOMEM;
+		err = vmap_pmd_range(pud, addr, next, phys_addr, prot, max_page_shift, mask);
+		if (err)
+			break;
 	} while (pud++, phys_addr += (next - addr), addr = next, addr != end);
-	return 0;
+	return err;
 }
 
 static int vmap_try_huge_p4d(p4d_t *p4d, unsigned long addr, unsigned long end,
@@ -268,6 +274,7 @@ static int vmap_p4d_range(pgd_t *pgd, unsigned long addr, unsigned long end,
 {
 	p4d_t *p4d;
 	unsigned long next;
+	int err;
 
 	p4d = p4d_alloc_track(&init_mm, pgd, addr, mask);
 	if (!p4d)
@@ -281,11 +288,11 @@ static int vmap_p4d_range(pgd_t *pgd, unsigned long addr, unsigned long end,
 			continue;
 		}
 
-		if (vmap_pud_range(p4d, addr, next, phys_addr, prot,
-					max_page_shift, mask))
-			return -ENOMEM;
+		err = vmap_pud_range(p4d, addr, next, phys_addr, prot, max_page_shift, mask);
+		if (err)
+			break;
 	} while (p4d++, phys_addr += (next - addr), addr = next, addr != end);
-	return 0;
+	return err;
 }
 
 static int vmap_range_noflush(unsigned long addr, unsigned long end,
-- 
2.43.5

Re: [PATCH v3] mm: vmalloc: WARN_ON if mapping size is not PAGE_SIZE aligned

[Uladzislau Rezki]

On Thu, Oct 09, 2025 at 05:37:06PM +0800, Yadong Qi wrote:
> In mm/vmalloc.c, the function vmap_pte_range() assumes that the
> mapping size is aligned to PAGE_SIZE. If this assumption is
> violated, the loop will become infinite because the termination
> condition (`addr != end`) will never be met. This can lead to
> overwriting other VA ranges and/or random pages physically follow
> the page table.
> 
> It's the caller's responsibility to ensure that the mapping size
> is aligned to PAGE_SIZE. However, the memory corruption is hard
> to root cause. To identify the programming error in the caller
> easier, check whether the mapping size is PAGE_SIZE aligned with
> WARN_ON().
> 
> Signed-off-by: Yadong Qi <yadong.qi@linux.alibaba.com>
> Reviewed-by: Huang Ying <ying.huang@linux.alibaba.com>
> ---
> v2 -> v3:
>   * change error code from ENOMEM to EINVAL
>   * modify callers of vmap_pte_range to handle return code
> v1 -> v2:
>   * Use WARN_ON instead of BUG_ON
> ---
>  mm/vmalloc.c | 29 ++++++++++++++++++-----------
>  1 file changed, 18 insertions(+), 11 deletions(-)
> 
> diff --git a/mm/vmalloc.c b/mm/vmalloc.c
> index 5edd536ba9d2..1fa52f203795 100644
> --- a/mm/vmalloc.c
> +++ b/mm/vmalloc.c
> @@ -100,6 +100,9 @@ static int vmap_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end,
>  	struct page *page;
>  	unsigned long size = PAGE_SIZE;
>  
> +	if (WARN_ON(!PAGE_ALIGNED(end - addr)))
>
And it might be worth to use WARN_ON_ONCE() otherwise there is a risk
that a kernel buffer would contain only such warnings.

--
Uladzislau Rezki

RE: [PATCH v3] mm: vmalloc: WARN_ON if mapping size is not PAGE_SIZE aligned

[yadong.qi]

> -----Original Message-----
> From: Uladzislau Rezki <urezki@gmail.com>
> Sent: 2025年10月9日 17:46
> To: Yadong Qi <yadong.qi@linux.alibaba.com>
> Cc: akpm@linux-foundation.org; urezki@gmail.com; linux-mm@kvack.org;
> linux-kernel@vger.kernel.org; ying.huang@linux.alibaba.com
> Subject: Re: [PATCH v3] mm: vmalloc: WARN_ON if mapping size is not PAGE_SIZE
> aligned
> 
> On Thu, Oct 09, 2025 at 05:37:06PM +0800, Yadong Qi wrote:
> > In mm/vmalloc.c, the function vmap_pte_range() assumes that the
> > mapping size is aligned to PAGE_SIZE. If this assumption is
> > violated, the loop will become infinite because the termination
> > condition (`addr != end`) will never be met. This can lead to
> > overwriting other VA ranges and/or random pages physically follow
> > the page table.
> >
> > It's the caller's responsibility to ensure that the mapping size
> > is aligned to PAGE_SIZE. However, the memory corruption is hard
> > to root cause. To identify the programming error in the caller
> > easier, check whether the mapping size is PAGE_SIZE aligned with
> > WARN_ON().
> >
> > Signed-off-by: Yadong Qi <yadong.qi@linux.alibaba.com>
> > Reviewed-by: Huang Ying <ying.huang@linux.alibaba.com>
> > ---
> > v2 -> v3:
> >   * change error code from ENOMEM to EINVAL
> >   * modify callers of vmap_pte_range to handle return code
> > v1 -> v2:
> >   * Use WARN_ON instead of BUG_ON
> > ---
> >  mm/vmalloc.c | 29 ++++++++++++++++++-----------
> >  1 file changed, 18 insertions(+), 11 deletions(-)
> >
> > diff --git a/mm/vmalloc.c b/mm/vmalloc.c
> > index 5edd536ba9d2..1fa52f203795 100644
> > --- a/mm/vmalloc.c
> > +++ b/mm/vmalloc.c
> > @@ -100,6 +100,9 @@ static int vmap_pte_range(pmd_t *pmd, unsigned long
> addr, unsigned long end,
> >  	struct page *page;
> >  	unsigned long size = PAGE_SIZE;
> >
> > +	if (WARN_ON(!PAGE_ALIGNED(end - addr)))
> >
> And it might be worth to use WARN_ON_ONCE() otherwise there is a risk
> that a kernel buffer would contain only such warnings.
> 

Sure, will change in next version.

Best Regard
Yadong